The purpose of this study is introduce of sampling method and prediction of hog population. Before computing sample size we cut off small farms from the sampling frame. The contribution from this part of the population is at least small in comparison with the remaining population. It may be tempting not to use resources on farms that contribute little to the overall result of the survey. Moreover, this reduces the response burden for these small farms. These farms were excluded from the frame so that the sample size could be reduced without significantly affecting quality. The sample design is based on 2000 agricultural census data and other related government data. Two stage stratified random sampling is applied for the sample design, where the first stage stratum is a nine geographical area(province), and the second stage is a seven stratum within province. The sample are 388 households for the pig.